Voice of the DBA

Tag: sql server

  • An Unexciting Exciting Release

    SQL Server 2025 was released this week. The announcement came at Ignite and the PASS Data Community Summit with keynotes on Wednesday and Thursday, respectively. While there are some things to look forward to in the release (What’s New) and some highlights from T-SQL Tuesday this week, it seems that this release wasn’t a very exciting one.

    On one hand, I blame all the Microsoft Fabric focus, which seems to distract from the core product that powers the databases at many organizations.  The SQL Server blog from Microsoft has had relatively few posts this year, highlighting a few things. The Fabric blog gets more posts, which is something I’ve seen at Database Weekly as well. As I curate the content during my week, I find a lot more Fabric-focused content than SQL Server-specific posts. That contributes to a lack of excitement for a new version of SQL Server.

    On the other hand, SQL Server is a very mature product and most of us use it daily, depend on it, and feel comfortable with the way it works. While some of us want changes, bugs fixed or features enhanced, for the most part, I find many clients expect it to just work and don’t have time to refactor code to use new features. A few also are nervous about using features until they’ve been in a few versions, as we sometimes see features released with limited functionality and few enhancements (*cough* these tables *cough* or these ones).

    At the same time, the What’s New page has a lot of stuff listed. If not more than SQL Server 2022 and 2019, it’s in more categories. Maybe it’s just that the documentation writers are more verbose? Perhaps, but I see lots of things I hadn’t noticed before in announcements. RegEx and the AI capabilities have dominated a lot of the news/blogs/etc., but the Change event streaming, native support for parquet/Delta, AG control flow, backup on secondaries, ADR for tempdb, and more are there. Plus, quite a few security enhancements to let us work with managed identities.

    When I really look at the product, there are things to be excited about, probably at least a couple for most people. At the same time, it has seemed like many people, including me, aren’t necessarily overly excited about upgrading systems to a new version when much of our code won’t change.

    Maybe because no one lets us, or gives us time to, change code to use new features.

    Steve Jones

    Listen to the podcast at Libsyn, Spotify, or iTunes.

    Note, podcasts are only available for a limited time online.

  • SQL Server 2025 GAs Today

    If you aren’t watching the Ignite keynotes today, then you might have missed the announcement that SQL Server 2025 officially releases today. The main SQL Server page gives you a place to get some marketing info, while the docs should change to remove “preview” today.

    I’ll have some thoughts later in the week, but this has been a low key release for me. While I’ve been hearing about different features, both publicly and under my MVP NDA, there hasn’t seemed to be a lot of people excited about the release.

    That makes some sense as the product is very mature, and most of us use the core features, which sometimes get improvements in new versions, and sometimes regressions.

    In any case, as I look through the What’s New, there is a lot of stuff changing, and a lot of additions. I always like the T-SQL language changes, and with so many people using JSON these days, having more functions to work with these strings is important. Regex is meh for me, but fuzzy string matching is interesting.

    SSMS v22 also goes GA today, and I hope it’s more stable than the preview has been. I’ve tried the GitHub Copilot in here, and I keep seeing freezes in the app. However, I also think this is a feature that will get a lot of data professionals experimenting with AI. Of course, if you have SQL Prompt, I’d suggest you try it there first.

    The big thing that I’d have appreciated in a lot of jobs, and I think a number of customers will like is a Standard Edition version of Developer Edition. That prevents mistakes where devs read about something on SQL Server Central or elsewhere and then implement it in dev, without realizing it won’t work in their production Standard Edition version.

    I don’t know how excited many of you are to use this version, but it’s the latest one and starts the clock on 10 years of support for security fixes. We’ve had a few lately, so at this point, if I were upgrading an older instance, SQL Server 2025 is where I’d go. Especially now that Standard Edition can run with 256GB of RAM. That alone might get me to upgrade some older SE instances.

  • T-SQL Tuesday #192: SQL Server 2025 Backup Changes


    I hosted this month, but I decided to put my own entry in as well. There are more things in this release than I expected, probably because of the overwhelming focus on Fabric from Microsoft that has drowned out other changes.

    I’ve been a DBA (or Ops manager) for quite a bit of my career, and one of the things I’ve thought was top priority was backups. I always want to be sure I have a recovery strategy, because if I don’t, nothing matters. Things go wrong and systems fail and I need to protect systems.

    Security is a close second, but backups really matter.

    I was excited to see that full and differential backups can be set on secondary replicas in AGs. To be fair, I had assumed this would be working by SQL Server 2022, but I hadn’t checked. Glad it’s been added. I know, I know, I questioned if you need HA, but that’s more a staffing question than one of whether HA makes sense in general.

    Backups, especially on larger systems, can be a load on the server. We’ve seen this with compressed backups natively, or with something like SQL Backup. As a result, offloading this to another system is something I’d want to do.

    Of course, there’s still the challenge of making sure you get these all organized in one place (don’t have the 1 copy of your diff only on one secondary), but that’s a separate post.

  • Your Favorite Improvement

    SQL Server 2025 is out in preview form. CTP 2.0 is available for you to test, and there are quite a few changes that have been added to the platform. Some of these are already in the cloud, but this is your chance to test them in your environment.

    This is version 17.x, though I don’t know we’ve really had 17 versions to work with. In any case, there have been a lot of versions in my career. I’ve connected to and worked on all of these: 4.2, 6.0, 6.5, 7.0, 2000, 2005, 2008, 2008R2, 2012, 2014, 2016, 2017, 2019, 2022. No 2025 yet (other than install and a few basic queries), but that’s coming.

    In that time, I’ve seen this platform evolve significantly, incorporating new features, capabilities, language changes, and more. SQL Server has been a very capable database platform for millions of users around the world. Many of you reading this have probably based a good portion of your career, if not all of it, on Microsoft SQL Server. I know I have, and it’s been a good career.

    With that in mind, and trying to be positive, what is your favorite feature/improvement/change in SQL Server over the years? What has made a big difference in your career or job? Maybe this made your customers happy, maybe something made your job easier, or maybe there’s a change that just reduced your stress.

    If you’re looking for a list of changes, there are some at Wikipedia, though this isn’t an exhaustive list. Brent wrote about his top 6 things, all of which are good changes. I certainly think that the DMVs are amazing when I look at other platforms and the lack of information they surface.

    If I think about the things that have made a huge difference for my positions, I’d say that 2005 with CTEs and DMVs were huge. However, for me, the addition of the OVER() clause and Window functions are incredible. That is one of the language changes I have used a lot over the years to simplify aggregate queries that were much harder with GROUP BY structures.

    The other feature that I think is amazing is the automatic seeding in Availability Groups. That alone makes it easy to add nodes, consider upgrades, and more without requiring long, slow backup copies. While I’ve rarely used it, I can think of many times in the past when it would have made my life much easier.

    What is your favorite improvement to the platform over the years? Let me know which thing has made your job much easier, more efficient, or maybe more enjoyable.

    Steve Jones

    Listen to the podcast at Libsyn, Spotify, or iTunes.

    Note, podcasts are only available for a limited time online.